Living Brain: shared memory for all your AI tools

What's new

On November 4, 2026, ChatAndBuild launched Living Brain, a persistent memory layer for AI. The memory doesn't live inside an individual chatbot but alongside it, so it can be stored and retrieved across ChatGPT, Claude, Cursor, Codex and users' own agents. The Product Hunt page also mentions Gemini, OpenClaw and Hermes. The company says the service works with more than 60 AI platforms, including ones from Anthropic, OpenAI, Google and NVIDIA. The day after launch, it ranked No. 2 on Product Hunt.

It can be accessed through a workspace, an API, an SDK (a toolkit for developers) and MCP (Model Context Protocol, a common standard that lets AI tools connect to external sources).

What's smart about it

You can start work in ChatGPT and continue in Claude or Codex without explaining the project from scratch. The connected tools can both read from and write to the shared memory, so it stays up to date.

Project updates can also be sent in through a ChatChat agent on WhatsApp, and they are then available in the connected tools. ChatChat is ChatAndBuild's platform for persistent AI agents. Content is assessed using five internal metrics, including Memory Worthiness and Re-entry Priority, and the state of the memory is tracked with Freshness, Strain and Readiness.

Cheaper or better?

The company says the system uses up to 22.6 times fewer tokens (chunks of text) per question than giving the AI the full history: 5.1 thousand versus 116 thousand on the LongMemEval-S test. On LoCoMo, the figure is 2.8 times fewer. The cost per task is said to be up to 95 percent lower. Accuracy is 92.2 percent on LoCoMo (1,540 questions, graded by an AI judge), and the company reports 0 percent hallucinations (fabricated answers) on LongMemEval-S and 94.7 percent on questions about time. Median retrieval time is 464 milliseconds, and end-to-end response time is 3.7 seconds.

All figures come from the company's own tests, and no independent verification is available. The company itself says the savings depend on workload and setup.

The pricing comes from FourWeekMBA and is not listed on the company's homepage. According to FourWeekMBA, the API costs $8 per million input tokens and $0.50 per million output tokens, with a minimum prepayment of $5. A workspace costs $99 per month with 3 team seats and 1 production project, and the Business tier has custom pricing. Whether a free version exists is not stated in the sources.

Living Brain is positioned as a complement to ChatGPT, Claude, Cursor and Codex, not a competitor. A comparison with other shared memory services is not provided in the sources.

Who it's good for

The service is for people who use several AI tools on the same project and are tired of explaining everything again each time. That includes teams and developers who build their own AI agents and want them to remember across conversations.

At $99 per month for a workspace, it is aimed first and foremost at teams and businesses rather than individuals. The sources don't say whether it is available in Denmark and the EU, or how data is handled.